Dynamic pointer trap-to-Date

Traditional indicators of Year-to-Date, Quarter-to-Date, Month-to-Date probably do not pose. There may be situations that require less traditional variation to-Date, which we call labor trap-to-Date, then the sum of all values ​​from the past until the selected date. Consider a table facts on bank loans, where one of the indicators is the amount of the loan - positive if the transaction increases the amount of the loan amount, while reducing the negative and zero if no impact on the amount of the loan. The transaction, terms of dimensions, we are interested, are recorded daily at the individual loan agreements.report required above the table should display změnu sum loan amount for each year (as of the year contributed to the total sum of the loans), while the absolute amount of all loans until the end of the year, including (the sum of all the changes the amount of loans from the past until the end of the year).








It is required to drillování to the level of individual days. Important for them to live above the loan contracts were always calculated from the infinite past - it disqualifies RSUM type functions, which allow only the returned result.


Elegant solution is to use the feature today, accompanied by adjusting the time dimension and its hierarchy.

  1. The time dimension is required to add a new column, which we call VSECHNY_DATUMY need and fill it the same value in all rows.
  2. We will adjust the time dimension hierarchy so that we add a new level between the Grand Total and Year and as a key column we give VSECHNY_DATUMY.
  3. Create the metadata pointer loan amount of all contracts using today as a parameter and enter the level Dummy.

Endpoint behaves in the way we used today. Allows freedom to drill the different levels of hierarchy time dimension, without requiring us to develop indicators somehow modified.